Suppose I measure reaction time of 4 students, initially before any treatment, then after treatment 1, and finally after treatment 2. (Note all treatments will eventually be applied to all students)
Also because of potential error with reaction time measurements, suppose 3 trials are done for each student-treatment combination.
Now if there was only 1 student-treatment combination, a simple repeated measures ANOVA could be applied.
In this case, is it statistically ok to instead use a randomized block design, where each block is comprised of the trials relating to a particular student?
